Real-Time Gender Classification Using MiniXception and Hand Gesture Detection Using MediaPipe Framework

In human-computer interaction, gender classification and hand gesture recognition are essential technologies. This paper presents a real-time system that integrates gender prediction and face and hand gesture detection, utilising the Mini Xception model and MediaPipe framework. We address the requirements for efficient real-time operation by utilizing frame-skipping mechanisms in our work. The system comprises video capture with OpenCV, model integration using TensorFlow, and efficient data handling with NumPy. Extensive testing has proven its generalization to diverse settings and excellent performance, with 84% accuracy for both tasks, capable of real-time processing.